Loss of Normal G₁ Checkpoint Control Is an Early Step in Carcinogenesis, Independent of p53 Status¹

It has long been known that mammalian cells can arrest in the G₁ phase of the cell cycle in response to ionizing radiation (1). However, it remains unclear whether there is a systematic difference between normal and neoplastic cells in the capacity to undergo such G₁ arrest, and what role, if any, G₁ checkpoint control may play in determining radiosensitivity or in the process of radiation-induced carcinogenesis.

A prolonged first cycle G₁ arrest after irradiation was initially described in normal human diploid fibroblasts (2). In experiments in which human diploid fibroblasts were synchronized in the G₀/G₁ phase before irradiation by contact inhibition or serum starvation and the progression of cells into S phase was monitored by continuous labeling with [³H]thymidine, a transient delay in the progression of cells from G₁ to S phase was observed, and a fraction of the cells appeared irreversibly arrested in G₁ (3). Both the G₁ delay and the prolonged G₁ arrest were completely abolished in human fibroblasts transfected with the E6 papilloma virus gene to abrogate p53 function (4, 5). In addition, E6-transfected cells were more resistant to cell killing by radiation (6, 7).

Interestingly, similar experiments with 14 human tumor cell lines demonstrated a much reduced capacity to undergo a G₁ arrest after irradiation compared to normal human diploid fibroblasts (8). G₁ arrest was completely absent in the tumor cell lines with mutant p53 status. The tumor lines expressing wt³p53 showed either no G₁ arrest or a minor arrest that was much smaller than that observed in human diploid fibroblasts (8). Similar results were obtained with wt p53 tumor cells synchronized in early G₁ phase by the mitotic shake-off technique (9) and with cells irradiated during exponential growth (10). These results suggested that tumor cells generally have a reduced capacity to undergo G₁ arrest after irradiation compared to normal cells, regardless of the p53 status of the tumor. This is consistent with the generally accepted concept that tumor cells have multiple genetic alterations, many of them involved in cell cycle control during the G₁ phase.

On the other hand, other laboratories have reported a prolonged p53-dependent first-cycle G₁ arrest after irradiation of some human tumor cell lines (11, 12, 13). However, most of these studies used nonsynchronized exponentially growing tumor cells and infrequent sampling at late times (16–20 h) after irradiation, raising the question of whether it was first- or second-cycle G₁ arrest that was measured (discussed in Ref. (9)). A pronounced first-cycle prolonged G₁ arrest has been reported in one human glioma cell line (14).

Whatever the controversy between previous reports, one weakness of studies with established human tumor cell lines is that many of them have been carried for long periods, often years, in culture. Tumor cells are known to be genetically unstable, and such long-term passage in vitro could result in the accumulation of additional genetic alterations. The reduced G₁ arrest observed in human tumor cell lines compared to early-passage normal human diploid fibroblasts could, thus, be due to a difference between late-passage established cell lines versus early-passage primary cells, rather than being a true difference between the normal and neoplastic state. Another complication with such a comparison is that most of the human tumor cells studied do not arise from fibroblasts, and they are also not derived from the same patient.

To clarify this issue and to explore the relationships between G₁ arrest, radiosensitivity, and genetic alterations in neoplastic cells, we have used a model system of normal and transformed mouse 10T1/2 fibroblast cell clones. Malignant transformation of individual cells is thought to represent an early stage in carcinogenesis. The advantage of this system is that the transformed cells can be directly compared with the normal cells from which they originated. The transformed cell clones developed over a 6-week period after irradiation, leading to minimal differences between the normal and transformed cells resulting from passaging alone. Furthermore, because radiation is used to induce transformation, the transformed clones are heterogeneous with respect to genetic alterations, which makes this a suitable system to study the effects of p53 status. Using this model system, we compared radiation-induced G₁ arrest in normal and transformed clones and explored the relationship between G₁ arrest, radiosensitivity and p53 status. All of the transformed clones examined showed the G₁ arrest to be absent or markedly reduced, suggesting that diminished G₁ checkpoint control is an early event in the process of carcinogenesis, independent of p53 status.

Cells from C3H 10T1/2 clone 8 mouse fibroblasts (15) were maintained at 37°C/5% CO₂ in Eagle’s basal medium supplemented with 10% heat-inactivated (56°C, 30 min) fetal bovine serum (Rehatuin, Intergen), 50 units/ml penicillin, and 50 μg/ml streptomycin.

Transformed clones of irradiated 10T1/2 cells were established as described previously (16). Briefly, confluent 10T1/2 cells (passage 8–10) were irradiated with 6 Gy of X-irradiation (220-kV and 15-mA X-rays; dose rate, 0.8 Gy/min) or 2-Gy α-irradiation (3.7-MeV α-particles from a ²³⁸Pu source; dose rate, 0.099 Gy/min) and immediately seeded at low densities to yield ∼300 viable cells per dish. The cells went through ∼13 rounds of cell division before the cultures entered the confluent, density-inhibited phase of growth. Culture medium was changed twice a week until cells became confluent and once a week thereafter. After 6 weeks, transformed clones (type III foci) overlying the confluent monolayer were isolated from the dishes and expanded in cell culture. Normal clones were isolated from nonirradiated 10T1/2 cells. All clones were tested for the ability to form colonies in soft agar.

A total of 10⁴ cells were seeded with 2 ml of culture medium to 10 dishes (diameter = 35 mm). On subsequent days thereafter, cells from one dish were trypsinized and the number of cells was counted using a Coulter particle counter (Coulter Limited Electronics, United Kingdom). Culture medium in the remaining dishes was changed twice a week. The exponential part of the growth curve was fitted to a straight line using the method of least squares, and cell doubling times were determined as ln2/k, where k is the slope. Saturation densities were obtained from the last points of the growth curves.

Plastic dishes (60 mm) were precoated with 5 ml of medium containing 0.5% agarose (type II; Sigma) and 20% fetal bovine serum. Cell suspensions at 10⁴ cells/1.5 ml were made in medium containing 0.2% agarose and 10% heat-inactivated fetal bovine serum, and 1.5 ml of this was spread over the precoated dish. Subsequently, cells were fed once a week with 1 ml of medium (without agarose). After 2 months, the number of colonies was scored by viewing the dishes through a microscope and counting four to eight separate areas marked on each dish. Pictures were taken with an Olympus PM20 camera.

G₁ arrest was measured by the cumulative labeling indices method (3, 8). Cells were synchronized in the G₀/G₁ phase before irradiation by holding cells at confluence and waiting 3 days after the last medium change. After exposure to 6 Gy of X-irradiation (160-kV and 18-mA X-rays; dose rate, 0.75 Gy/min), cells were immediately subcultured at low density with medium containing 1 μCi/ml [³H]thymidine (20 Ci/mmol; New England Nuclear, Boston, MA). Dishes of irradiated and control cells were fixed in 100% ethanol at appropriate times thereafter. Standard autoradiographic techniques were used; 200 cells were scored on each dish to determine the cumulative labeling indices.

Confluent cells were irradiated (X-ray) and immediately suspended and reseeded at low density to measure colony-forming ability. Duplicates with two different cell numbers were plated for each radiation dose. After 14 days, cells were fixed in 70% ethanol, stained with crystal violet, and counted. Colonies containing more than 50 cells were scored as survivors.

Confluent cells were irradiated with 6 Gy (X-rays), and samples were collected 3 h later. Briefly, cells were pelleted and lysed in immunoprecipitation assay buffer [50 mm Tris-Cl (pH 7.5), 150 mm NaCl, 1% NP40, 0.5% sodium deoxycholate, 0.1%SDS, 1 mm phenylmethylsulfonyl fluoride, and 5 mm EDTA]. After centrifugation, protein quantification was carried out by the Bradford method (Bio-Rad). The protease inhibitors aprotinin, leupeptin, and pepstatin were added (1 μg/μl). Seventy μg of protein were loaded to each lane in 12% polyacrylamide gels. The antibodies used were anti-p53 (Ab-7) and anti-p21^Waf1 (Ab-6) from Oncogene Science.

DNA extraction was performed using a modified version of the salting-out method (17). Briefly, cells were pelleted and resuspended in 300 μl of lysis buffer (10 mm Tris-HCl, 400 mm NaCl, 2 mm EDTA, pH 8.0) supplemented with 30 μl of 10% SDS and 15 μl of 10 mg/ml proteinase K. After overnight digestion at 37°C, 150 μl of saturated NaCl was added. Samples were agitated for 15 s and centrifuged at 13000 rpm for 10 min. The supernatant containing DNA was transferred to a new tube, and DNA was precipitated by 2.5 volumes of ethanol.

The presence of mutations in the mouse p53 gene was determined by PCR-SSCP analysis and subsequent sequencing. Exons 5–8 were amplified by PCR using the following primer set (Refs. 18 and 19; purchased from Life Technologies, Inc.): exon 5, 5′-GAC ACC TGA TCG TTA CTC GG-3′ (upstream) and 5′-GGA GGC TGC CAG TCC TAA CC-3′ (downstream); exon 6, 5′-GGT TAG GAC TGG CAG CCT CC-3′ (upstream) and 5′-GTC AAC TGT CTC TAA GAC GC-3′ (downstream); exon 7, 5′-GAC TTC ACC TGG ATC CTG TG-3′ (upstream) and 5′-CTA ACC TAA CCT ACC ACG CG-3′ (downstream); and exon 8, 5′-TTC TTA CTG CCT TGT GCT GG-3′ (upstream) and 5′-TGA AGC TCA ACA GGC TCC TC-3′ (downstream). The PCRs were performed in a 40-μl volume containing: 500 ng of sample DNA; 50 pmol of upstream and downstream primer; 6.3 μCi of [³⁵S]dATP (1250 Ci/mmol; New England Nuclear, Boston, MA); 4 units of AmpliTaq DNA polymerase (Perkin-Elmer/Cetus, Norwalk, CT); 1.25 μmol of dATP; 12.5 μmol of dCTP, dGTP, and dTTP; 2 mm MgCl₂; and 4 μl of 10× PCR buffer II (Perkin-Elmer). The PCR amplification was performed using a programmable thermal controller (MJ Research Inc.) with 35 cycles of 1 min at 94°C, 1 min at 57°C, and 1 min at 72°C, and a final step of 10 min at 72°C. After PCR, 5 μl of the final PCR product were mixed with 5 μl of SSCP dilution buffer (0.1% SDS-10 mm EDTA) and 10 μl of Stop Solution (United States Biochemical; 95% formamide, 20 mm EDTA, 0.05% bromphenol blue, and 0.05% xylene cyanol). Samples were heated to 90°C for 3–4 min and analyzed on a 6% polyacrylamide gel with 10% glycerol. The gel was run at room temperature for 16 h at 6 W and exposed to film for 2–4 days.

DNA sequencing was performed using the Sequenase PCR Product Sequencing Kit (United States Biochemical) with the same primer set as that used for the PCR-SSCP analysis. The products were analyzed on a 6% polyacrylamide denatured sequencing gel. The gel was run at 1500 V for 3–5 h, dried, and exposed to film for 3–5 days for autoradiography.

Cells from the normal 10T1/2 clones formed a flat, density-inhibited monolayer at confluence and were not able to grow in soft agar (Fig. 1 and Table 1), characteristic of a normal phenotype. The cell doubling times of the normal clones were ∼20 h. In contrast, cells from transformed clones that were established from type III foci of irradiated 10T1/2 cells grew to higher densities at confluence and formed colonies in soft agar (Fig. 1 and Table 1), consistent with a transformed phenotype. It is interesting to note the large variations between individual transformed clones in both the cloning efficiency and size and shape of colonies in soft agar as well as the saturation densities and cell doubling times (Fig. 1 and Table 1). This heterogeneity may be due to differences in genetic alterations between the transformed clones.

One of the clones, clone 6, which originally was established from a type III focus, failed to form colonies in soft agar and grew to a saturation density similar to that of normal clones (Fig. 1 and Table 1). Clone 6 may thus have reverted back to the normal phenotype, as has been described previously (20). This clone was included in our study as a normal control clone, which differs from the other normal clones in that it has survived a radiation dose similar to the transformed clones.

Nonirradiated normal 10T1/2 clones showed low levels of p53 and p21 protein, which were increased 2.5–5.0- and 1.5–3.0-fold, respectively, at 3 h after a dose of 6 Gy of irradiation (Fig. 2), consistent with a normal, functional p53 gene. Similarly, nine of the transformed cell clones showed a functional p53 gene, with increased p53 and p21 protein levels after exposure to 6 Gy (examples shown in Fig. 2). The remaining three transformed clones (clones 7, 15, and 17), however, showed very high levels of p53 protein and no regulation of p53 or p21 levels after irradiation (example shown in Fig. 2), suggesting that p53 was mutant in these clones. This was confirmed by PCR-SSCP analysis of exons 5–8 of the p53 gene and subsequent sequencing. The same point mutation with a single base shift from C→G at codon 132 (exon 5) was found in clones 7 and 17, and a point mutation with a single base shift from G→C at codon 135 (exon 5) was found in clone 15 (Fig. 3). No evidence of mutation was found by sequencing of the parental wt cells.

To accurately measure first cycle G₁ arrest in the normal and transformed cell clones, cells were synchronized in G₀/G₁ by confluence holding and subsequent subculture to low density. The cumulative progression of cells into the S phase was monitored by uptake of [³H]thymidine. Fig. 4 shows the results from representative experiments. Parental 10T1/2 cells showed a transient G₁ delay lasting ∼9 h after exposure to 6 Gy, as measured by the delay time at 50% of the maximum cumulative labeling index (Fig. 4,A). Transformed clone 9 with wt p53 status showed a G₁ delay of ∼3 h (Fig. 4,B), and transformed clone 7 with mutant p53 status showed no G₁ delay (Fig. 4,C). The results of similar measurements of G₁ delay in the whole panel of normal and transformed clones are summarized in Fig. 5. All of the normal clones showed a transient G₁ delay lasting ∼9 h after 6 Gy of irradiation. In comparison, G₁ delay was markedly reduced in the transformed clones. The three transformed clones with mutant p53 and two of the clones with wt p53 lacked an X-ray induced G₁ arrest, whereas the remaining transformed clones showed short delays, again with some variation among the individual clones.

Standard clonogenic survival assays were performed after irradiation. Survival curves derived from the mean survival levels at each radiation dose for normal and transformed clones are shown in Fig. 6. Although the mean survival levels for transformed clones were slightly higher, no significant difference between normal and transformed clones is evident, in agreement with two previous reports (21, 22). The Ps were 0.17 (2 Gy), 0.25 (4 Gy), and 0.34 (6 Gy). The results for the individual clones are presented in Fig. 7. Interestingly, the three transformed clones with mutant p53 were no more radioresistant than the clones expressing wt p53. To better assess any eventual relationship between G₁ arrest and radiosensitivity, G₁ delay times were plotted as a function of radiosensitivity (Fig. 8). From Fig. 8, it is clear that there was no relationship between G₁ arrest and radiosensitivity in this model system of normal and transformed mouse cells.

The results of this study indicate that the radiation-induced first-cycle G₁ arrest is markedly diminished in transformed as compared to normal mouse 10T1/2 cells, regardless of p53 status. Because the transformed cells arise from the same normal cells in this system, they come from the same genetic and histology background as the normal clones. In light of a model for tumor development involving multiple steps during the progression from an initiation event to a fully malignant invasive tumor (23), the changes we have observed most likely correspond to the time around tumor initiation. The results, thus, strongly support the conclusion that loss of normal G₁ checkpoint control may be an early event in tumor development, independent of p53 status. This finding is consistent with previous reports of a diminished G₁ arrest after irradiation in wt p53 human tumor cell lines (8, 9, 10). Tumor cells may thus be defective in G₁ cell cycle control in response to ionizing radiation as well as to a variety of growth control signals (24, 25).

Radiation-induced G₁ arrest is known to be mediated, at least in part, by p53-dependent transcriptional activation of the cyclin-dependent kinase inhibitor p21 (26, 27). Nine of the transformed clones in our study showed a functional p53 gene in terms of p53 and p21 protein induction. The lack of a normal G₁ arrest in these clones indicates that p21 may not be functioning properly, perhaps due to other genetic alterations that occurred during transformation. It is of interest in this context that p21 appeared to be constitutively overexpressed in the transformed as compared with normal clones, and that its induction by radiation was weaker (Fig. 2).⁴ Elevated constitutive levels of p21 protein have also been described in many wt p53 human tumors (28). It is not known how this might relate to the reduced capacity for a G₁ arrest. Clearly, further studies will be needed to determine the exact mechanisms by which transformed cells expressing wt p53 show a reduced G₁ arrest; however, these may vary among individual clones because they likely have different genetic alterations.

The results presented in Fig. 8 showed no relationship between G₁ arrest and radiosensitivity in transformed mouse cells. Thus, the extent of G₁ arrest alone cannot determine radiosensitivity, although it may play a role in some cases, such as in human normal fibroblasts (6, 7). Furthermore, we found no relationship between p53 status and radiosensitivity in the transformed mouse cells. This is in agreement with previous reports from our group and others in which no relationship between p53 status and radiosensitivity was found in most human tumor cell lines (8, 29, 30). Other studies have reported that some tumors with mutant p53 tend to be more radioresistant (31, 32, 33). The lack of a strict relationship between p53 status and radiosensitivity in tumor cells may be because of multiple genetic alterations, some of which may lead to enhanced radioresistance, whereas others enhance radiosensitivity. In contrast, there is a clear relationship between abrogating p53 function by E6 transfection and enhanced radioresistance in normal human fibroblasts where no other genetic alterations are present (6, 7). This is consistent with preliminary findings in p53−/− mouse embryonic fibroblasts.⁵

We found p53 mutations in 3 of 12 (25%) radiation-induced transformed clones, consistent with a previous screening where p53 mutations were found in 31% of transformed clones (16). Interestingly, two of the clones had the same point mutation at codon 132; the third clone had a point mutation very close by at codon 135. Both these mutations are located in a part of the DNA-binding region of the p53 gene that is highly conserved among species (34). In fact, Cys-132 in the mouse p53 gene is a homologue to Cys-135 in the human p53 gene (34), which is known as a relative hotspot for mutations in human cancer (35). Thus, radiation-induced p53-mutations in the 10T1/2 mouse model system may be highly relevant with respect to p53-mutations in human cancer. One important issue to address for future studies with this model system is at what time these p53 mutations occur and whether they are likely to be a cause or a consequence of the transformation process. Apparently, two events must have occurred during the 6 week period transformed foci developed: mutation of one allele and loss of heterozygosity at the other.

In conclusion, we have shown that radiation-induced G₁ arrest is reduced in transformed as compared to normal mouse 10 T1/2 cells, regardless of p53 status. We also found no relationship between G₁ arrest and radiosensitivity or p53 status and radiosensitivity in this model system. These results suggest that diminished G₁ checkpoint control may be an early event in the genesis of a tumor, one that is independent of p53 status.

We thank Drs. Edouard I. Azzam, Hatsumi Nagasawa, and Yongjia Yu for expert technical advice and for helpful discussions.